Blog Viewer

From web analytics practitioner to Chief Data Officer. What could be the career path?

By Alban Gerome posted 12-04-2018 06:31 PM

  
Good evening all,

It has been a few months since I last posted here. I am based in London where I have spent most of my career as either a web developer or a web analytics practitioner. I have now over 6 years' experience in database, server-side and front-end development but more recently, an additional 10 years in web analytics, mostly on the implementation side of things. Over the past year or so I began reading articles about the emerging role of Chief Data Officer with great interest. In May this year, I met Peter Jackson, not the filmmaker but the CDO at local water utility company near London as both of us presented at the Data Festival London. Peter was about to start a free online course to explain to us with another fellow CDO what the role means, how according to Gartner the companies will start needing CDOs en masse from next year onwards. Peter Jackson and Caroline Carruthers were the teachers and based the course on the book they co-wrote, The CDO Playbook, which has great reviews and which I also recommend. I am proud to say that I have completed that course as a student of the very first class of the Chief Data Officer School. But I am none the wiser about the what the career path toward becoming a CDO could be like. Well, almost...

I was the only web analytics practitioner in that class. Peter and Caroline considered that people like me can become a CDO but it would be much harder than people with different data backgrounds. More recently, I was checking the website of a UK-based recruiter specialised in Analytics called Harnham. I spotted a new blogging section on their website and CDO, Noam Zeigerson, was sharing his thoughts there. Now Harnham does not recruit CDOs so I was rather intrigued to see a CDO posting there so I reached out to Noam and asked him for his thoughts about what this elusive career path could look like. Although Noam did not have any clear answers, he did suggest that web analytics is seen as different to any other kind of data-related career. First, the type of data we handle is different from any other kind of data the organisations generate or acquire. Second, the kind of tools we use to analyse web analytics data are limited to web analytics data and do not let us analyse any other kind of data.

One possible career path might be becoming a Head of Analytics first. I lead a discussion at MeasureCamp Brussels last October. I did very basic research ahead of the discussion by looking at the LinkedIn profiles of several Head of Analytics. I was only able to look at 20 profiles before LinkedIn blocked me from viewing more profiles so my sample is tiny to be sure. But it suggested that only 3 in 20 Head of Analytics had any prior web analytics experience. The most typical profile was someone with a Masters in either Statistics, Maths, Economics or an MBA with some kind of data science experience. I guess it makes sense. The rare few Head of Analytics with a web analytics background suggest that the role title combines really two very different roles and the web analytics flavoured Head of Analytics seems to be the rarer of the two. This suggests that, at least from where I sit which is the United Kingdom, there are very opportunities to move ahead in web analytics and other data careers once you had 10 years' experience in the field. Many of us here are becoming contractors in our quest to earn a bigger paycheck as a result.

Another possible path might come from leveraging our favourite web analytics tools' APIs along with R and Python. At this stage, we only use the web analytics web interface when it is expedient. Most of the time, we extract the web analytics data, clean this data, combine it with other data sources. After all, I am used to seeing more and more presentations at MeasureCamp in Europe about R, Machine Learning with TensorFlow, Markov Chains-driven attribution modelling, Sentiment Analysis, Kubernetes and Docker. At this point, it seems to me the typical user of our web analytics web interfaces is less and less a web analyst but more and more one of your organisations' colleagues from all other teams. Perhaps we are beginning to see a convergence of the two types of Head of Analytics and even the emergence of a hybrid of both types. These hybrid Head of Analytics might have one foot in the web analytics realm and another one in the pure data science realm.

Web analytics data is certainly a lot messier than any other type of data that organisations see. It is rarely normally distributed, you only get if the customers gave their consent and if they support Javascript and cookies. I believe that these hybrid data-scientists are more likely to come from the ranks of web analysts experimenting with R and Python than from data scientists wondering what is web analytics is all about. After all, their lack of exposure to web analytics does not seem to limit their chances to become a Head of Analytics as I explained above. With greater visibility, the hybrid data scientist might appear as a stronger choice for organisations looking for a Head of Analytics and even more senior roles. As a web analytics implementation expert, perhaps I should learn R and Python, a few sophisticated approaches in these languages to analyse data from other sources than web analytics and pursue more data science roles such as Associate Data Architect and eventually a CDO perhaps. Please let me your thoughts. How is it like over the pond? What are the typical career paths of people with 10 years' experience in web analytics? Can web analytics people have a better chance senior data roles such as CDOs?

Alban
@albangerome



#CareerAdvice

Permalink

Comments

12-13-2018 03:30 PM

Ian for your comment. It's absolutely amazing to read that you have been on that career journey yourself and made it, found my post and took the time to comment back and share with us your experience! I totally agree with you about expanding beyond web analytics and into more data-related fields. I am looking into AWS at the moment with great interest.

There was this great article on the Harvard Business Review by Cassie Kozyrkov earlier this month. She is the Chief Decision Scientist at Google and she was arguing that analysts are now getting a raw deal compared to machine learning and AI people. But without analysts, these other types of data people have nothing to dig their teeth into. This resonated with what I thought I was seeing in the profiles of Head of Analytics people, i.e. suspiciously few people with analytics experience.

Well, I could not help but post a lengthy but positive comment on Cassie's article. 3 points mainly:

  1. If an analyst does not show an appetite for learning the business, it's a bad sign according to Cassie. I think it takes two to tango. If my recommendations as an analyst were systematically ignored I would stop bothering after a while and try my luck elsewhere. If, on the contrary, some of my early recommendations got implemented I might develop this appetite and a pretty voracious one too, hooked on the feeling of getting stuff done
  2. Cassie believes that when analyst finds something odd in the data, they should submit it to a business stakeholder next before a statistician or a machine learning engineer starts looking into it. I think that the analysts, statisticians, machine learning and even the implementation guys, should be part of a single analytics centre of excellence and interact with the rest of the business as in the hub and spoke model. Only when the finding has gone through all the roles of the analytics team should it reach the business. Starting the discussion with the business early is probably ok, perhaps to establish the potential business value but not necessarily in all cases. This makes me think of machine learning in unsupervised mode
  3. Most of the time though, I believe that an analytics team should operate in a way similar to supervised mode instead. A business stakeholder should start with a business question. Then the analytics team selects data sources that might answer the question or get the implementation guy to deploy new code and start collecting some or acquire data from third-parties. The team cleans the data, make hypotheses, run tests. This Plan-Do-Check-Act goes through a few iterations until an answer comes out which they submit back to the business stakeholder

Anyway, I digressed somewhat from my original topic but Cassie has got our analysts backs in her article. Ian confirmed that we should expand into other data-related roles such as learning R or Python, stats machine learning so with AWS I might be on the right track.

12-06-2018 07:24 PM

Hi Alban,

Thanks for this interesting post. As someone from a web analytics background who has just taken a new job as a CDO, I may have some encouragement for you :) Web Analytics is not the only foundation you need to be a CDO, but if you broaden the definition slightly to digital or event analytics, then it is a very important foundation. Most CDO roles today are in organisations with a strong digital presence and much of their data is generated from digital sources, so web analytics concepts are a good grounding.

That said, in my view a CDO needs to be able to have a view over a number of related areas to do with data and analytics:

  • Data Engineering & technology
  • Digital Analytics
  • Executive Reporting & Dashboarding
  • Data Science & Statistics
  • Data Strategy
In my own journey to CDO, I started out in web analytics (which bundles a number of the above disciplines together and minimizes others, which is I think one of the reasons that some data folk think that it is a "poor man's" data discipline) and then did tours of duty in various of the others - for example, for three years I ran a team that was entirely devoted to data engineering and technology, which taught me a ton of useful stuff about how to build sustainable and scalable data platforms. More recently my role has focused much more on advanced analytics and Data Science (even though I am not a Data Scientist myself), and in the conversations I've been having, this breadth has been very useful.

Hope that's useful!

Ian

Most Recent Blogs

Log in to see this information

Either the content you're seeking doesn't exist or it requires proper authentication before viewing.